The centerpiece of the work is a tightly controlled feeding experiment. Participants consumed an identical standardized carbohydrate meal on separate occasions, once alone and once in the company of a person to whom they felt close. Continuous glucose monitoring revealed that the same food produced a different metabolic story depending on the social context. When participants ate in companionship, their post-meal glucose excursions were smaller, meaning the spike in blood sugar after eating was blunted, and their glucose levels returned to baseline more quickly. In physiological terms, the body needed to deploy less of its regulatory machinery to keep a potentially destabilizing nutrient load under control. Because repeated large glucose swings are implicated in the long arc of metabolic disease, the finding has immediate relevance for one of the most pressing public health challenges of the era.

Crucially, the researchers did not treat this single result as an isolated curiosity. Across a series of preregistered experiments, they tested whether the same proximity effect would appear in other homeostatic domains. It did. When adults were exposed to cold, those who were near a close social partner maintained their temperature regulation more effectively than those tested alone. The pattern also emerged in the earliest stages of life: infants, whose survival depends entirely on the bodies of their caregivers, showed more efficient stress regulation when held or kept in close contact. The consistency of the effect across glucose control, thermoregulation, and stress responsivity is what elevates the work from an interesting observation to a candidate general principle, one that cuts across organ systems that are usually studied in isolation from one another.

Two features of the findings are particularly striking from a technical standpoint. First, the benefits of companionship appeared even without physical touch. Mere proximity to a loved individual was sufficient, suggesting that the relevant signal is not tactile contact or shared heat but something subtler, likely a learned association in which the presence of a specific person predicts safety and reduced physiological demand. Second, the researchers found that the effect was not explained by reductions in subjective stress. This is a critical distinction, because a simpler interpretation of the data would be that companions merely feel calmer, and that calmness secondarily improves metabolic readings. The evidence points instead to a direct regulatory benefit, as if the body recalibrates its operating costs downward when it detects a reliable ally nearby, independent of how stressed the person reports feeling.

Prof. Atzil framed the discovery as a challenge to one of the deepest assumptions in biomedical science. We used to think of the body as an isolated system that regulates itself, she noted, but the findings show that human physiology is inherently social. In her account, being close to someone you love lowers the physical energy the body spends on self-regulation, producing a direct metabolic dividend for togetherness. This framing carries an evolutionary implication that the researchers make explicit: if bodies genuinely run better in company, then social bonding may not be merely a strategy for protection, cooperation, or reproduction layered on top of individual physiology. It may be woven into the very economics of homeostasis, giving organisms a concrete energetic reason to form and maintain attachments in the first place. Human bonds, on this view, are partly metabolic instruments.

The mechanistic story that emerges is one of energetic accounting. Keeping internal conditions stable, a process biologists call homeostasis, is not free. Every correction the body makes, secreting insulin to clear glucose from the bloodstream, shivering or redirecting circulation to defend core temperature, releasing and then dampening stress hormones, consumes energy and imposes wear on tissues. If the nervous system can verify, through continuous monitoring of the social environment, that a protective partner is present, the allostatic load, the cumulative cost of staying ready for threat, can be lowered. The parasympathetic branch of the autonomic nervous system, which favors restorative states, can gain greater influence, and predictive circuits in the brain can downshift their threat forecasts. The result is a body that achieves the same regulatory targets with less expenditure, which is precisely the signature the experiments detected.

The developmental dimension of the work deserves particular emphasis. Infants cannot regulate their own temperature, glucose, or stress responses competently on their own; for most of human evolutionary history, an infant separated from caregivers faced rapid physiological collapse. The finding that infant stress regulation is more efficient in close proximity is therefore unsurprising in outline, but it gains new meaning when read alongside the adult data. It suggests that the social modulation of physiology does not disappear with maturity, as many researchers assumed, but persists throughout life as a quiet background process. The adult eating lunch with a friend and the newborn cradled against a parent may be running variations on the same ancient regulatory program, one in which the presence of a bonded other is counted as an external resource that the body can draw upon to lighten its internal workload.

The public health implications arrive at a moment when dysregulated blood sugar has become a global concern. Rates of type 2 diabetes and prediabetes continue to climb across income levels and continents, and standard advice centers on diet composition, portion control, physical activity, and medication. This study does not overturn any of that guidance, and the researchers are careful to state that social closeness is not a substitute for medical care, nutrition, or exercise. But it does introduce a variable that almost no clinical guideline currently addresses: the identity of the people sitting at the table. If identical meals produce meaningfully different glycemic responses depending on companionship, then epidemiological models of metabolic health that ignore relational context may be systematically missing part of the picture. Blood sugar regulation, as Atzil puts it, is not just about what we eat but who we eat with, and relationships may hold an active biological role in physical wellbeing.

The research design strengthens the credibility of these claims in ways worth noting for readers weighing how much to trust a single headline. The experiments were preregistered, meaning the researchers committed to their hypotheses and analysis plans before collecting data, a safeguard against the selective reporting that has troubled social science in the past. The meal challenge was standardized, removing the obvious confound that shared meals might simply involve different foods. The replication of the effect across distinct physiological domains, and across age groups from infancy to adulthood, makes it unlikely that the result is an artifact of one particular task or population. At the same time, open questions remain. The precise neural and hormonal pathways that translate the perception of a companion into altered insulin dynamics or thermal efficiency have not yet been fully mapped, and future work will need to identify whether oxytocinergic signaling, vagal tone, or other mechanisms carry the signal.

What the study ultimately offers is a revised map of what a human body is. For more than a century, physiology has been practiced largely as the science of the sealed individual, a self-contained machine measured in isolation within laboratory walls. The concept of Social Physiology proposes instead that the boundary of the regulating system extends beyond the skin, into the network of trusted others whose presence the nervous system tracks at all times. Human connection, long celebrated in poetry and studied in psychology as a source of meaning and happiness, now appears in the laboratory as something harder-edged: a measurable reduction in the biological cost of staying alive. The people beside us, the researchers conclude, constitute a powerful and overlooked factor in health, one that operates not through inspiration or encouragement but through the quiet, continuous recalibration of glucose curves, thermal defenses, and stress responses. Relationships, it turns out, do not merely support our emotional lives. They help run the machinery of the body itself.

Subject of Research: The effect of social proximity on human glucose, temperature, and stress regulation

Article Title: Eating with a loved one lowers your blood sugar

Article References: Eating with a loved one lowers your blood sugar. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: social physiology, blood glucose, homeostasis, stress regulation, thermoregulation, metabolic health, Science Advances, Hebrew University of Jerusalem, social bonding, preregistered experiments, infant development, public health

Interoception, the central concept in the new analysis, is commonly defined as the process by which the nervous system senses, interprets, and integrates signals arising from within the body. Heartbeat, breathing rhythm, temperature, visceral tension, and the constellation of internal sensations that color every waking moment all travel upward through dedicated neural channels. The authors emphasize that a substantial portion of this traffic passes through the spinal cord, forming what they call spinal interoceptive pathways. These pathways deliver continuous updates about bodily states to a distributed interoceptive system in the brain, where they inform and constrain predictive models of what the body will experience next. In this framework, the brain is not a passive receiver but an active forecaster, constantly generating predictions about upcoming bodily states and comparing them against incoming signals.

The predictive dimension is where interoception connects most directly to emotion. According to the theoretical models reviewed by Romo-Nava and his colleagues, the brain uses its predictions about bodily states to adjust physiology through descending projections, a closed-loop architecture that maintains homeostasis—the finely tuned internal equilibrium on which survival depends. Emotional experience, in this view, is inseparable from the quality and accuracy of the bodily signals feeding the loop. When the ascending traffic is distorted, dampened, or amplified, the brain’s predictive models drift out of alignment with actual bodily conditions, and the subjective result may be the persistent malaise, heaviness, and dysphoria characteristic of depression. The authors argue that dysregulated signaling in spinal interoceptive pathways and the resulting interoceptive processing errors may play a genuine role in the depressive syndrome rather than being a mere byproduct of it.

What makes this proposal more than a philosophical reframing is its therapeutic implication. If the spinal cord carries a continuous stream of body-to-brain information that shapes mood, then the spinal cord is a physically accessible target. Unlike dispersed cortical networks buried deep within the brain, spinal pathways run through an anatomically well-mapped structure that clinicians already know how to stimulate, record from, and modulate. Neuromodulation technologies—devices and techniques that alter nerve activity through targeted electrical or other stimulation—are routinely applied to the spinal cord in other areas of medicine, most famously for pain management. The perspective article argues that the same general strategy could, in principle, be redirected toward the interoceptive traffic that informs emotional state, opening a novel route for studying and eventually treating major depressive disorder.

The scale of the unmet need gives the proposal urgency. Major depressive disorder affects hundreds of millions of people worldwide, and a large fraction of patients do not achieve lasting relief from available medications and psychotherapies. Existing neuromodulation treatments for depression, such as transcranial magnetic stimulation and implanted brain stimulation devices, work from the top of the loop, targeting the brain directly. An approach that intervenes from the bottom of the loop, at the spinal gateway where bodily signals enter, would represent a fundamentally different mechanistic strategy—one that could complement rather than compete with brain-focused therapies. The authors position spinal interoceptive modulation not as a replacement for existing care but as an additional lever on a system that current treatments leave largely untouched.

Yet the researchers are candid about a major obstacle: a substantial knowledge gap. Despite the theoretical appeal of the framework, the role of spinal interoceptive pathways in major depressive disorder remains poorly characterized. It is not yet established which specific spinal mechanisms are altered in depression, how those alterations propagate to the distributed interoceptive system in the brain, or how changes in spinal signaling correlate with changes in mood. The perspective article frames this gap as the central scientific challenge of the field and argues that closing it requires deliberate experimentation rather than assumption. The authors present their work as emerging evidence supporting the exploration of spinal interoceptive pathways as novel and feasible targets—emphasizing both words, novelty and feasibility, to signal that the approach is grounded in existing technology while pointing toward unexplored biology.

A key methodological problem the article confronts is how to prove that a spinal intervention is actually doing what it claims to do. In clinical neuroscience, this is the problem of target engagement: demonstrating that a treatment measurably modifies the biological mechanism it is designed to influence. Without reliable markers of spinal interoceptive target engagement, any clinical trial of spinal neuromodulation for depression would be flying blind, unable to distinguish between a therapy that failed because the mechanism was wrong and one that failed because the delivery was imprecise. The authors therefore devote significant attention to candidate experimental models and the critical next steps needed to identify markers that would show whether a given intervention is genuinely changing spinal interoceptive signaling. Such markers would be the measurement backbone for future studies, allowing researchers to connect stimulation parameters to physiological changes and physiological changes to clinical outcomes.

The research program described in the article is already backed by institutional investment. The work was supported by the Lindner Center of Hope and the University of Cincinnati, as well as by the National Institute of Mental Health, which funded the effort in part through grant 1R61MH133770-01A1. The breadth of the author team reflects the interdisciplinary demands of the project: the collaboration includes psychiatrists, neurologists, neurosurgeons, a biostatistician, and interoception specialists from the Laureate Institute for Brain Research and UCLA’s Semel Institute for Neuroscience and Human Behavior. Bringing together expertise in spine-level intervention, brain imaging, psychiatric assessment, and trial design is a prerequisite for a program that spans the length of the neuraxis—from peripheral bodily sensors through spinal relay stations to cortical interoceptive networks.

The broader scientific context makes the timing of this proposal notable. Across neuroscience, interoception has moved from a niche topic to a central framework for understanding psychiatric illness, with conditions including anxiety, eating disorders, and depression increasingly interpreted as disorders of bodily signal processing. Predictive processing models of the brain have given the field a common theoretical language, describing perception and emotion as consequences of prediction and prediction error rather than simple stimulus response. The spinal cord, in this emerging picture, is no longer a mere cable between brain and body but an active processing station that shapes the information it transmits. The Romo-Nava perspective extends that shift by insisting that the clinical implications run in both directions: just as the brain constrains the body through descending projections, the body constrains the brain through ascending spinal channels, and either side of the loop can be an intervention point.

If the research program succeeds, the consequences could extend well beyond depression. Spinal interoceptive pathways plausibly contribute to the physical symptoms that accompany many psychiatric conditions—chronic pain, fatigue, autonomic dysregulation—and a validated method for measuring and modulating spinal interoceptive signaling could illuminate all of them. For now, the authors’ contribution is a roadmap: a reasoned case that the spinal cord is a feasible, mechanistically motivated target for depression research, an honest account of what remains unknown, and a set of candidate models and engagement markers to guide the experiments ahead. As the field confronts the limits of treatments aimed solely at the brain, the message of this perspective is that mood may be built, in part, from signals that begin their journey in the body—and that the gateway through which those signals travel may one day become a place where depression is treated.

Subject of Research: The role of spinal interoceptive pathways in major depressive disorder and their potential as neuromodulation targets

Article Title: Spinal interoceptive pathways as therapeutic targets in depression

Article References: Romo-Nava, F., Awosika, O. O., Phan, P., Basu, I., Liu, J. C., Charnas, C., Georgiev, G., Mori, N. N., Welge, J., Blom, T., Fleck, D. E., Cao, X., Khalsa, S., Paulus, M., & McElroy, S. L. (2026). Spinal interoceptive pathways as therapeutic targets in depression. Discover Mental Health. https://doi.org/10.1007/s44192-026-00597-z

Image Credits: AI Generated

DOI: 10.1007/s44192-026-00597-z

Keywords: interoception, spinal cord, major depressive disorder, neuromodulation, predictive processing, homeostasis, psychiatry, spinal interoceptive pathways, target engagement, neuroscience, Spinal, interoceptive

The Gini coefficient takes a value between zero and one. In economics, a value of zero would mean everyone earns exactly the same income, while a value approaching one means a single individual holds nearly all the wealth. The researchers reasoned that the same logic applies to metabolite concentrations measured across a population of samples. If a molecule is homeostatically regulated by cells, tissues, or the organism as a whole, its concentration should be similar from person to person, yielding a low Gini coefficient. Conversely, a molecule that is exogenous, such as a pharmaceutical drug that only some individuals have ingested, should show wildly unequal concentrations across a cohort, producing a Gini coefficient close to one.

To test this idea, the team mined publicly available metabolomics datasets, including a large study of more than 200 identified plasma metabolites measured in 1,125 individuals with chronic obstructive pulmonary disease, available through the Metabolomics Workbench. The results were striking. Endogenous metabolites, which the authors call endogenites, peaked in their Gini distribution at around 0.2, while exogenous molecules such as drugs and food-derived compounds peaked above 0.95. The median Gini coefficient across the entire dataset was 0.263. When the researchers classified molecules by origin, 73 percent of exogenous molecules had Gini coefficients above 0.5, compared with just 2.9 percent of molecules considered endogenous or regulated.

Among the most tightly controlled molecules were the amino acids. Essential and non-essential amino acids had identical average Gini coefficients of just 0.14, and nine of the 25 lowest Gini values in the dataset belonged to amino acids including methionine, arginine, proline, serine, phenylalanine, asparagine, tryptophan, lysine, and glutamine. Glutamine itself recorded the lowest value of all, a Gini coefficient of 0.066 with a 95 percent confidence interval of just 0.063 to 0.068, a figure even lower than any observed in the team’s earlier transcriptomics analyses. Given that glutamine is a major hub of nitrogen metabolism, such extreme uniformity makes biological sense and suggests the molecule could even serve as a normalisation standard in metabolomics studies where sample volumes are uncertain.

At the opposite extreme sat pharmaceutical compounds. The anticonvulsant lamotrigine, for example, posted a Gini coefficient of 0.99, meaning its presence in plasma was almost entirely confined to the small subset of participants taking the drug. In a second dataset of 681 serum metabolites from 340 individuals studied in the context of tuberculosis, the highest values belonged to metabolites of paracetamol and aspirin. The researchers also examined vitamins, which occupy an interesting middle ground: they are essential and therefore physiologically important, yet exogenous in origin. Their Gini coefficients fell in an intermediate range of roughly 0.2 to 0.4, consistent with partial regulation as cofactors, though the values varied more than threefold across vitamins, likely reflecting differences in diet, absorption, supplementation, and microbiome interactions.

The study’s second focus was ergothioneine, a sulfur-containing amino acid derivative with the formula C9H15N3O2S and an exact monoisotopic mass of 229.0885 Da. Humans cannot synthesise this compound; it comes entirely from the diet, most notably mushrooms, and is transported into tissues by a dedicated transporter. Growing evidence links higher ergothioneine levels to reduced risks of cardiovascular disease, cognitive decline, dementia, and frailty, and previous work by the same group showed that women with high plasma ergothioneine were far less likely to develop pre-eclampsia. Because ergothioneine is exogenous but clearly physiologically important, the team predicted it would show an intermediate Gini coefficient, and the data confirmed this. Across multiple independent studies, ergothioneine’s Gini coefficient clustered consistently between 0.3 and 0.4: 0.38 in the COPD dataset, 0.373 in the tuberculosis cohort, 0.325 in a whole-blood dementia study, 0.457 in an ageing study, and 0.37 to 0.4 in a large dementia cohort from Singapore.

The analytical chemistry behind these measurements is itself noteworthy. Ergothioneine’s protonated form has a mass-to-charge ratio of 230.0958 in positive electrospray ionisation mode, and no other biologically relevant molecule lies within even 10 parts per million of this value, making database searches for the compound unusually straightforward. In new experimental work reported in the paper, the team measured ergothioneine in 40 antenatal serum samples from a pilot study at Liverpool Women’s Hospital, using ultra-high performance liquid chromatography coupled to an Orbitrap Exploris 240 mass spectrometer at a resolution of 120,000, with calibration solutions spanning 0.01 to 500 micromolar.

The Liverpool pilot delivered two surprises. First, the median ergothioneine concentration was just 180 nanograms per millilitre, far below the 261 nanograms per millilitre median seen in the earlier SCOPE study of 432 pregnant women; in fact, 180 nanograms per millilitre corresponds only to the ninth percentile of the earlier cohort. Second, women who went on to develop pre-eclampsia did not show the expected lower ergothioneine levels. The authors suggest this apparent contradiction dissolves once the population’s very low baseline is recognised: when nearly everyone is deficient, the protective relationship with concentration is obscured. Intriguingly, the Gini coefficient in the Liverpool cohort was lower than in almost all other ergothioneine studies, hinting that a depressed Gini value, even without absolute concentrations, might flag a population with inadequate ergothioneine exposure and a likely need for supplementation.

The researchers propose rough interpretive thresholds: metabolites with Gini coefficients below about 0.25 are subject to significant homeostasis or show low variation in exogenous supply, while those above about 0.75 are likely exogenous and largely unregulated. Molecules in between, like most vitamins and the nutraceuticals ergothioneine and kynurenic acid, are probably exogenous but partially regulated. The authors caution that a high Gini coefficient could sometimes reflect analytical error, missing values, or variable pharmacokinetics, making the metric best viewed as hypothesis-generating. A low value, however, is hard to explain away, and reliably indicates tight biological control. The team also notes that urinary metabolomes did not show systematically higher Gini coefficients than plasma, and that applying the approach to gut microbiome-derived metabolites awaits raw data that are not yet publicly available.

Beyond its technical contribution, the work carries a broader message about health inequality. The Liverpool findings, with median ergothioneine levels sitting at the ninth percentile of a comparable cohort, echo documented patterns of socioeconomic disparity in British health and mortality statistics. If a simple statistic borrowed from economics can simultaneously identify which molecules the body defends, expose hidden dietary deficits, and strengthen the case for targeted nutritional intervention, the Gini coefficient may prove to be one of the most versatile imports metabolomics has ever received from the social sciences.

Subject of Research: Using the Gini coefficient as a surrogate measure of metabolite regulability and homeostasis, with a focus on the diet-derived antioxidant ergothioneine

Article Title: The Gini coefficient as a surrogate for the regulability or homeostasis of metabolite concentrations: focus on ergothioneine

Article References: Kell, D. B., Dunn, W. B., Winder, C. L., Anand, K., Greenfield, B., Kenny, L. C., Merriel, A., Moore, J. B., & Waitt, C. (2026). The Gini coefficient as a surrogate for the regulability or homeostasis of metabolite concentrations: focus on ergothioneine. Metabolomics, 22(5), Article 151. https://doi.org/10.1007/s11306-026-02534-1

Image Credits: AI Generated

DOI: 10.1007/s11306-026-02534-1

Keywords: Gini coefficient, metabolomics, ergothioneine, homeostasis, metabolite regulation, nutraceutical, pre-eclampsia, mass spectrometry, amino acids, vitamins, dietary antioxidants, health inequality

MASLD is characterized by excess fat in liver cells in association with metabolic risk factors such as obesity, insulin resistance or type 2 diabetes. It can progress from relatively simple steatosis to metabolic dysfunction-associated steatohepatitis, fibrosis, cirrhosis and liver cancer. The disease is usually discussed in relation to diet, body weight, glucose regulation and inherited susceptibility, but environmental exposures may modify these processes. Bisphenols are endocrine-disrupting chemicals used in plastics, food and beverage packaging, thermal paper and other materials. BPA is the best-known member of the group, while bisphenol F and bisphenol S are among the substitutes increasingly used in products marketed as BPA-free. Biomonitoring studies have detected BPA and its analogues in human urine and blood, indicating widespread exposure. In the United States, estimates based on earlier national survey data placed median daily BPA intake at roughly 30 to 70 nanograms per kilogram of body weight. Exposure patterns are changing: restrictions have reduced BPA in some regions, while use of replacement compounds such as BPS and BPF has increased. The review argues that chemical substitution does not automatically remove biological concerns, because structurally related compounds can interact with hormone-sensitive pathways.

Iron is essential for oxygen transport, mitochondrial energy production and many enzyme reactions, but both too little and too much can damage the liver. The hormone hepcidin acts as the principal systemic regulator. Produced mainly by the liver, hepcidin binds to the iron-export protein ferroportin and causes it to be internalized and degraded. This reduces iron release from intestinal cells, macrophages and storage cells into the bloodstream. The review describes how iron deficiency can impair mitochondrial electron transport and fatty-acid oxidation, causing reactive oxygen species to rise while fat accumulates inside liver cells. Deficiency also reduces the activity of iron-dependent antioxidant systems and can stabilize hypoxia-inducible factors, or HIF proteins. HIF-1α and HIF-2α then alter gene programs controlling glucose and lipid metabolism. In particular, HIF-2α may increase expression of lipogenic genes such as SCD-1, FASN and ACC while suppressing PPARα and CPT1A, key regulators of fatty-acid breakdown. These changes can promote the buildup of triglycerides, diacylglycerol, ceramides and other lipids that interfere with insulin signaling and fuel inflammation.

Iron excess creates a different but equally damaging route to liver injury. Ferrous iron can drive the Fenton reaction, generating highly reactive hydroxyl radicals that attack proteins, membranes and mitochondrial structures. The resulting lipid peroxidation produces compounds such as malondialdehyde and 4-hydroxynonenal, which can further inhibit enzymes required for fatty-acid oxidation. Oxidative stress also activates inflammatory pathways including NF-κB, MAPK and JNK, while weakening PPARα activity and depleting glutathione, a major cellular antioxidant. If glutathione peroxidase 4, or GPX4, is impaired, lipid peroxides can accumulate to levels that trigger ferroptosis, a form of regulated cell death driven by iron-dependent membrane damage. Dying hepatocytes can activate Kupffer cells, the liver’s resident macrophages, which release inflammatory and fibrogenic signals such as tumor necrosis factor alpha and transforming growth factor beta. Those signals stimulate hepatic stellate cells, the principal producers of scar tissue in the injured liver. The review therefore portrays iron overload not as a passive marker of MASLD, but as a potential amplifier of steatosis, hepatocyte death and fibrosis.

Bisphenols may connect these two iron-related states through hormone receptors and redox signaling. Evidence summarized in the review points particularly to the G protein-coupled estrogen receptor, or GPER, as a candidate mediator of BPA-induced liver toxicity. Activation of GPER may stimulate an EGFR–PI3K–AKT–mTORC1 signaling cascade, suppressing NRF2-dependent antioxidant defenses and reducing glutathione synthesis. At the same time, bisphenols may increase reactive oxygen species by disturbing mitochondria and activating NADPH oxidase. Loss of glutathione can inactivate GPX4, allowing lipid peroxides to accumulate. The review proposes that this oxidative environment may then increase expression of iron-import proteins such as transferrin receptor 1 and divalent metal transporter 1, drawing more iron into hepatocytes. The additional iron would intensify Fenton chemistry and create a self-reinforcing cycle of iron accumulation, oxidative injury and impaired lipid metabolism. Animal and cell studies support parts of this model, including links between BPA or BPS exposure, altered iron-regulatory genes, ferroptosis and hepatic fat accumulation. However, the authors stress that the complete sequence has not been demonstrated in exposed human livers.

Hepcidin may help explain the review’s proposed paradox of hepatic iron overload alongside systemic iron deficiency. Bisphenols could affect hepcidin differently in different tissues or at different stages of exposure. In the liver, inflammation induced by chemical stress may raise interleukin-6 and activate the JAK2–STAT3 pathway, increasing hepcidin transcription. Oxidative stress may also stimulate NRF2, which can influence hepcidin-related regulatory elements. Excess hepcidin would reduce ferroportin, trapping iron in hepatocytes and macrophages. At the same time, the estrogen-like activity of bisphenols may act through estrogen receptor alpha and potentially suppress hepcidin expression, based partly on evidence from estrogenic regulation of iron metabolism and related environmental chemicals. Lower systemic hepcidin would stabilize ferroportin in intestinal cells, increasing iron absorption and altering the distribution of iron between the gut, blood and liver. The balance between these opposing signals may depend on dose, exposure duration, chemical structure and receptor abundance in each tissue. The authors call this proposed mismatch “hepcidin dyssynchrony,” but identify it as a working hypothesis rather than a settled mechanism.

The intestine adds another layer to the possible interaction. Iron availability strongly shapes the gut microbiome. During iron deficiency, increased intestinal uptake can leave less iron for microbes, favoring organisms able to produce siderophores, molecules that capture iron efficiently. During iron excess, oxidative chemistry may harm sensitive beneficial bacteria while allowing more stress-resistant pathobionts to expand. In either setting, the review describes possible reductions in bacteria associated with beneficial metabolites and increases in members of the Enterobacteriaceae family. Changes in short-chain fatty acids, including acetate, propionate and butyrate, may weaken signals that normally activate AMPK, support fatty-acid oxidation and restrain SREBP1c-driven lipogenesis. Lower butyrate may also reduce support for intestinal tight-junction proteins such as ZO-1 and occludin. A leakier intestinal barrier could permit more lipopolysaccharide to reach the liver through the portal circulation, where it activates Toll-like receptor 4 and NF-κB inflammatory signaling. Altered bile-acid metabolism may compound the effect by disturbing the FXR–FGF19/15 pathway, a key regulator of lipid and bile-acid homeostasis. These links remain biologically plausible, but the direction of cause and effect between iron imbalance, microbiome changes and MASLD is not yet clear.

The evidence base has important limitations that shape what can reasonably be concluded. Some preclinical experiments cited in the review used BPA doses of about 5 milligrams per kilogram per day, far above estimated median population exposure. Such studies can reveal molecular pathways and potential adverse outcomes, but their results cannot be transferred directly to human risk. Low-dose endocrine-disrupting chemicals may also produce non-monotonic dose responses, meaning that biological effects do not necessarily increase in a simple straight line as exposure rises. The review notes that the CLARITY-BPA program reported biological effects at several lower doses, although interpretation of low-dose findings remains an active scientific debate. Human studies have mainly examined associations between urinary or circulating bisphenol measurements and liver or metabolic outcomes, which cannot by themselves prove causation. Exposure is also usually mixed, involving multiple bisphenols, dietary factors, air pollutants and medications. The authors recommend population cohorts that measure bisphenol metabolites together with hepcidin, ferroportin-related markers, iron stores, liver fat and inflammatory indicators over time. They also call for physiologically based pharmacokinetic modeling, stable-isotope iron tracing, multi-omics analysis and studies using controlled environmental concentrations.

If the proposed pathway is confirmed, it could broaden strategies for preventing or treating environmentally influenced MASLD. Possible research directions include blocking inappropriate GPER signaling, restoring antioxidant capacity, correcting iron distribution or modifying gut microbial metabolism. Iron chelation or supplementation would require particular caution because both deficiency and excess can worsen metabolic injury, and treatment would need to be guided by reliable measures of tissue and systemic iron rather than by a single blood value. Microbiome-directed approaches might focus on preserving short-chain-fatty-acid production and intestinal barrier function, but they remain experimental for this application. For now, the review’s main contribution is conceptual: it places iron homeostasis at the intersection of endocrine disruption, redox biology, lipid metabolism and the gut–liver axis. The authors do not claim that bisphenols alone explain MASLD, a disease with many interacting causes. Instead, they suggest that common environmental chemicals may act as metabolic modifiers whose effects are amplified by obesity, insulin resistance, diet or pre-existing inflammation. Establishing whether this mechanism operates at real-world exposure levels will require direct human evidence, but the proposed framework offers specific molecular and clinical markers that future studies can test.

Subject of Research: Bisphenol-related disruption of iron homeostasis in metabolic dysfunction-associated steatotic liver disease

Article Title: Bisphenols perturb systemic iron homeostasis and fuel the pathogenesis of metabolic dysfunction-associated steatotic liver disease

Article References: Xu, J., Xu, H., Qi, R., Li, Y., Zhou, Z., Wu, W., Tian, Z., & Tang, Z. (2026). Bisphenols perturb systemic iron homeostasis and fuel the pathogenesis of metabolic dysfunction-associated steatotic liver disease. Discover Toxicology, 3(1), Article 19. https://doi.org/10.1007/s44339-026-00065-x

Image Credits: AI Generated

DOI: 10.1007/s44339-026-00065-x

Keywords: Bisphenols, Iron homeostasis, MASLD, Hepcidin, Ferroptosis, Oxidative stress, Gut microbiome, Liver disease, perturb, systemic, iron, homeostasis